> ## Documentation Index > Fetch the complete documentation index at: https://docs.getdialed.ai/llms.txt > Use this file to discover all available pages before exploring further. # Change a batch's priority or pacing > Re-order or re-pace a batch that is still queued or in flight. Both fields are optional — send only what you are changing, and an empty body leaves the batch untouched. - `priority` re-orders the batch's REMAINING work within its own dispatch lane. Work already claimed or sent is never re-ordered, and the boost cannot take capacity from another lane — it only changes the order in which your own account's budget is spent, which is why no elevated role is required. - `pacing` sets or replaces the release policy and applies from the next release tick: slowing a batch down never recalls work already released, and speeding one up never produces a burst to catch up. A policy that is not strictly slower than the rate your provider budget already allows is rejected rather than quietly clamped, so the policy you set is always the policy in force. Send `remove_pacing: true` to stop pacing a batch entirely. Explicit JSON `null` is rejected — omit a field to leave it unchanged. A completed, failed, or cancelled batch cannot be changed (409). Cross-tenant access returns 404, never 403. ## OpenAPI ````yaml /openapi.json patch /flows/batches/{batch_id} openapi: 3.1.0 info: title: GetDialed API version: 0.1.0 servers: - url: https://api.getdialed.ai/v1 description: Production security: [] paths: /flows/batches/{batch_id}: patch: tags: - batches summary: Change a batch's priority or pacing description: >- Re-order or re-pace a batch that is still queued or in flight. Both fields are optional — send only what you are changing, and an empty body leaves the batch untouched. - `priority` re-orders the batch's REMAINING work within its own dispatch lane. Work already claimed or sent is never re-ordered, and the boost cannot take capacity from another lane — it only changes the order in which your own account's budget is spent, which is why no elevated role is required. - `pacing` sets or replaces the release policy and applies from the next release tick: slowing a batch down never recalls work already released, and speeding one up never produces a burst to catch up. A policy that is not strictly slower than the rate your provider budget already allows is rejected rather than quietly clamped, so the policy you set is always the policy in force. Send `remove_pacing: true` to stop pacing a batch entirely. Explicit JSON `null` is rejected — omit a field to leave it unchanged. A completed, failed, or cancelled batch cannot be changed (409). Cross-tenant access returns 404, never 403. operationId: patch_batch_flows_batches__batch_id__patch parameters: - name: batch_id in: path required: true schema: type: string title: Batch Id requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/BatchPatchIn' responses: '200': description: Successful Response content: application/json: schema: $ref: '#/components/schemas/BatchResponse' '401': description: Missing or invalid authentication content: application/json: example: detail: Authentication required '404': description: >- No such batch for this organization (cross-tenant access also returns 404) content: application/json: example: detail: Batch not found '409': description: >- The batch is in a terminal state, so there is nothing left to re-order or re-pace content: application/json: example: detail: >- Cannot change a completed, failed, or cancelled batch — a terminal batch has nothing left to re-order or re-pace. '422': description: >- The body failed validation — an explicit null on a patchable field, an unknown priority, a policy that is not strictly slower than the effective provider budget, or `pacing` and `remove_pacing` sent together content: application/json: example: detail: pacing may not be null — omit it to leave it unchanged. '429': description: Rate limit exceeded — retry after the Retry-After header headers: Retry-After: description: Seconds to wait before retrying schema: type: integer content: application/json: example: detail: 'Rate limit exceeded: 100 per 1 minute' security: - APIKeyHeader: [] - HTTPBearer: [] components: schemas: BatchPatchIn: properties: priority: anyOf: - type: string enum: - normal - high - type: 'null' title: Priority pacing: anyOf: - $ref: '#/components/schemas/BatchPacingPolicyIn' - type: 'null' remove_pacing: type: boolean title: Remove Pacing default: false type: object title: BatchPatchIn description: >- Change a batch that is still queued or in flight. `priority` re-orders the batch's remaining work within its own dispatch lane; work already sent is untouched. `pacing` sets or replaces the release policy and applies from the next release tick — slowing a batch down never recalls work already released, and speeding one up never produces a burst to make up the difference. To stop pacing a batch entirely send `remove_pacing: true` rather than a null `pacing`. Every field is optional; send only what you are changing. example: pacing: period_minutes: 15 records_per_period: 5000 priority: high BatchResponse: properties: id: type: string title: Id definition_id: type: string title: Definition Id org_id: type: string title: Org Id trigger_id: type: string title: Trigger Id trigger_type: type: string title: Trigger Type trigger_source: anyOf: - type: string - type: 'null' title: Trigger Source schedule_id: anyOf: - type: string - type: 'null' title: Schedule Id metadata: additionalProperties: true type: object title: Metadata status: type: string enum: - pending - queued - scheduled - running - paused - completed - completed_with_errors - partially_failed - failed - cancelled title: Status record_count: type: integer title: Record Count scheduled_at: anyOf: - type: string format: date-time - type: 'null' title: Scheduled At started_at: anyOf: - type: string format: date-time - type: 'null' title: Started At completed_at: anyOf: - type: string format: date-time - type: 'null' title: Completed At created_at: type: string format: date-time title: Created At updated_at: type: string format: date-time title: Updated At priority: anyOf: - type: string enum: - normal - high - type: 'null' title: Priority pacing: anyOf: - $ref: '#/components/schemas/PacingPolicy' - type: 'null' type: object required: - id - definition_id - org_id - trigger_id - trigger_type - metadata - status - record_count - scheduled_at - started_at - completed_at - created_at - updated_at title: BatchResponse example: completed_at: '2026-07-09T18:00:05Z' created_at: '2026-07-09T17:59:59Z' definition_id: def_a1b2c3d4 id: batch_55667788 metadata: campaign: spring-2026 org_id: org_a1b2c3d4 record_count: 2 started_at: '2026-07-09T18:00:00Z' status: completed trigger_id: trigger_44556677 trigger_type: api_call updated_at: '2026-07-09T18:00:05Z' BatchPacingPolicyIn: properties: records_per_period: type: integer maximum: 1000000 minimum: 1 title: Records Per Period period_minutes: type: integer maximum: 1440 minimum: 1 title: Period Minutes window: anyOf: - $ref: '#/components/schemas/BatchDeliveryWindowIn' - type: 'null' type: object required: - records_per_period - period_minutes title: BatchPacingPolicyIn description: >- A ceiling on how fast a batch is released: N records every M minutes. The rate is a hard ceiling over any interval — there is no catch-up burst. After a pause, an exhausted provider quota or a closed delivery window, delivery resumes at the configured rate from that moment and the lost time is not made up. `period_minutes` is a whole number of minutes; pacing finer than a minute cannot be expressed. Add a `window` to confine delivery to particular local hours. Released records are ordinary queue work: they may still be merged with other batches' work bound for the same destination, but this batch never delivers faster than its policy allows. example: period_minutes: 15 records_per_period: 5000 window: days: - monday - tuesday - wednesday - thursday - friday end_minute_of_day: 1020 start_minute_of_day: 540 timezone: America/Denver PacingPolicy: properties: records_per_period: type: integer minimum: 1 title: Records Per Period period_minutes: type: integer minimum: 1 title: Period Minutes window: anyOf: - $ref: '#/components/schemas/DeliveryWindow' - type: 'null' type: object required: - records_per_period - period_minutes title: PacingPolicy description: >- A ceiling on how fast a batch may be released: N records every M minutes. The rate is a hard ceiling over any interval — there is no catch-up burst. After a pause, an exhausted provider quota or a closed delivery window, delivery resumes at the configured rate from that moment; the time lost is not made up. `period_minutes` is measured in whole minutes: pacing finer than a minute cannot be expressed. Add a `window` to confine delivery to particular local hours. BatchDeliveryWindowIn: properties: timezone: type: string minLength: 1 title: Timezone start_minute_of_day: type: integer exclusiveMaximum: 1440 minimum: 0 title: Start Minute Of Day end_minute_of_day: type: integer maximum: 1440 minimum: 1 title: End Minute Of Day days: anyOf: - items: type: string type: array - type: 'null' title: Days type: object required: - timezone - start_minute_of_day - end_minute_of_day title: BatchDeliveryWindowIn description: >- The hours of the day during which a paced batch may deliver. `timezone` is a required IANA name such as `America/Denver` — a fixed UTC offset is wrong for half the year, so a window always states its own zone and there is no default. `start_minute_of_day` and `end_minute_of_day` are minutes since local midnight, start inclusive and end exclusive; `end_minute_of_day` may be `1440` (local midnight ending the day). A window that would wrap past midnight is refused. `days` limits delivery to particular lowercase weekday names (`monday` … `sunday`); omit it for every day. Records still queued when the window closes are neither dropped nor rushed: delivery resumes at the same local start time on the next allowed day. example: days: - monday - tuesday - wednesday - thursday - friday end_minute_of_day: 1020 start_minute_of_day: 540 timezone: America/Denver DeliveryWindow: properties: timezone: type: string title: Timezone start_minute_of_day: type: integer exclusiveMaximum: 1440 minimum: 0 title: Start Minute Of Day end_minute_of_day: type: integer maximum: 1440 minimum: 1 title: End Minute Of Day days: anyOf: - items: type: string type: array - type: 'null' title: Days type: object required: - timezone - start_minute_of_day - end_minute_of_day title: DeliveryWindow description: >- The hours of the day during which a paced batch may deliver. `timezone` is a required IANA name such as `America/Denver` — a fixed UTC offset is wrong for half the year, so there is no default and no fallback: a window always states its own zone. `start_minute_of_day` and `end_minute_of_day` are minutes since local midnight, start inclusive and end exclusive; `end_minute_of_day` may be `1440` (local midnight ending the day). A window that would wrap past midnight is rejected. `days` limits the window to particular lowercase weekday names (`monday` … `sunday`); omit it for every day. Records left over when the window closes are not dropped and are not rushed: delivery resumes at the same local start time on the next allowed day. securitySchemes: APIKeyHeader: type: apiKey in: header name: X-API-Key HTTPBearer: type: http scheme: bearer bearerFormat: JWT ````